Installation

Requirements

The NLP Architect requires Python 3.6+ running on a Linux* or UNIX-based OS (like Mac OS). We recommend using the library with Ubuntu 16.04+. We recommand installing basic OS compilers and python development packages.

Before installing the library make sure you has the most recent packages listed below:

Ubuntu* 16.04+ or CentOS* 7.4+ Mac OS X* Description
python-pip pip Tool to install Python dependencies
python-dev python-dev Python development dependencies
libhdf5-dev h5py Enables loading of hdf5 formats
pkg-config pkg-config Retrieves information about installed libraries

Note

The installation of NLP Architect will install CPU-based binaries of all deep learning frameworks. For specific installation of backends of Tensorflow or PyTorch (CPU/MKL/GPU) we recommend installing NLP Architect and then installing the desired backend DL framework.

Prerequisites (creating virtual env)

Make sure pip and setuptools and venv are up to date before installing.

pip3 install -U pip setuptools venv

Note

We recommend installing NLP Architect in a virtual environment to self-contain the work done using the library. Users can use any virtual environment (pip, conda, etc.)

To create and activate a new virtual environment (or skip this step and use the wizard below):

python3.6 -m venv <my_new_env>
source <my_new_env>/bin/activate

Install from pip

pip install nlp-architect

Install from source

git clone https://github.com/NervanaSystems/nlp-architect.git
cd nlp-architect
pip install -e .

Running Examples and Solutions

To run provided examples and solutions please install the library with [all] flag which will install extra packages required. (requires installation from source)

pip install .[all]

Updating NLP Architect

Depending of how you installed NLP Architect to update the library:

Using pip

pip install -U nlp-architect

From source

git pull origin master